Predicting which cancer patients will best respond to therapy is an enormous health care issue. It has been recently suggested, based on early reads of whole exome sequencing data, that there may only be ~12 molecular pathways that drive the development and progression of cancer. Many experimental therapies are now targeting these pathways. We believe that gene expression signatures may be one of the best ways to judge the activation of a particular molecular pathway, by providing a molecular summary of the activity of many genes. Moreover, we believe that the selective addition of mutational assessment may improve the resolving power of a hybrid, multi-analyte (DNA + RNA) test that may be used to guide patients to the most effective drugs. We have recently developed gene expression signatures to measure the activation of two of the most important pathways in colon cancer, RAS and PI3K, for which there is an increasing availability of pathway targeted therapeutics. Due to the complex nature of these pathways, simple analysis of canonical single gene mutations only identifies the response characteristics of a proportion (<30%) of the population. Here, we propose to technically validate the existing RAS and PI3K signatures and to refine their activity through novel mutational assessment. Multi-analyte signatures/ algorithms will be clinically validated in a CLIA environment with a cohort of colorectal cancer patients treated with cetuximab therapy. This approach will prepare signatures for clinical application in the near future.